Arsenite efflux limits microbial arsenic volatilization CURRENT STATUS: REVIEW

Background: Arsenic (As) methylation is regarded as a potential way to volatize and thereby remove As from the environment. However, most microorganisms conducting As methylation display low As volatilization efficiency as As methylation is limited by As efflux transporters as both processes compete for arsenite As(III). In the study, we deleted arsB and acr3 from Rhodopseudomonas palustris CGA009, a good model organism for studying As detoxification, and further investigated the effect of As(III) efflux transporters on As methylation. Results: Two mutants were obtained by gene deletion. Compared to the growth inhibition rate (IC50) [1.57±0.11 mmol/L As(III) and 2.67±0.04 mmol/L arsenate As(V) of wildtype R. palustris CGA009, the As(III) and As(V) resistance of the mutants decreased, and IC50 value of the R. palustris CGA009 ∆arsB mutant was 1.47±0.02 mmol/L As(III) and 2.12±0.03 mmol/L As(V), respectively, and that of the R. palustris CGA009 ∆acr3 mutant was 1.21±0.07 mmol/L As(III) and 1.76±0.12 mmol/L As(V), respectively. The As volatilization rate of R. palustris CGA009 ∆arsB and R. palustris CGA009 ∆acr3 was 7.36 and 10.46 times higher than that of the R. palustris CGA009 at 100.0 µmol/L As(III) when incubated for 12 h, respectively, and 7.21 and 10.30 times higher than that of the R. palustris CGA009 when incubated with 25.0 mol/L As(V), respectively. At 25.0 mol/L As(III), low doses of methylarsonate MAs(V) and dimethylarsonate DMAs(V) were detected in both the wild type and in the deletion mutant content of medium the indicating that Acr3 the highest As(III) efflux rate. Conclusion: The results of this study showed that As efflux transporters were shown to be a remarkable intrinsic factor limiting As volatilization efficiency, and As volatilization rate could be significantly improved by deleting genes encoding microbial As efflux transporters. Our study provided

an explanation for the often low rate of microbial As methylation and an effective strategy for screening microorganisms with high As volatilization. Background Arsenic (As) is a highly toxic metalloid widely distributed in nature, and is considered to be one of the toxic compounds that affect human health and environmental contamination that are of great concern [ 1 ]. The transport and transformation of As in the environment are governed by geochemical as well as biological processes, generating an As biogeochemical cycle [ 2 ]. Microorganisms play a key role in driving the As biogeological cycle [ 3 ]. Microbes have developed multiple strategies to handle arsenic including As resistance (ars) systems, which reduces cytoplasmic arsenate [As(V)] to arsenite [As(III)] which is then transported across the cytoplasmic membrane [ 4 , 5 ]; the As(III) oxidation (aio) system, which oxidizes As(III) to As(V) [ 4 , 6 ]; the As(V) respiration (arr) system, which respire and reduce As(V) to As(III) [ 4 , 7 ]; and the As methylation (arsM) system, which converts inorganic arsenic into methylated As compound [ 8 ]. These strategies inspired people to employ these As-metabolizing microbes to remove As from contaminated sites, especially As volatilization that is dependent on As methylation. As methylation catalyzed by As(III) S-adenosylmethionine methyltransferase (ArsM) converts highly toxic As(III) into low-toxic pentavalent mono-, di-and trimethylarsenic compounds, and finally forms volatile trimethylarsenic [TMAs(III)] under aerobic conditions [8][9][10]. Numerous investigations have shown evidence of As volatilization in different environments such as landfill, peatland, paddy soil, geothermal sites and sewage treatment sites [8,[11][12][13]. Therefore, As methylation and subsequent volatilization can be exploited as a potentially effective measure for bioremediation of As contamination.
As methylation is widespread in the natural environment and is a prerequisite for the production of volatile methylarsine gases [ 14 ]. Many microbes, both wild type and engineered strains, are able to methylate As and subsequently volatilize As. However, As volatilization, which is dependent on As methylation, is not efficient for most microorganisms except for Arsenicibacter rosenii SM-1 (47.6±18.4%) [9], possibly Prochlorococcus [ 15,16] and engineered Pseudomonas putida KT2440 (31%) [17]. On the one hand, some environmental parameters such as organic matter [13,18] 22 ] are able to affect As volatilization efficiency. These parameters might alter the expression level and activity of the gene products of microbial arsMs, which are the genes encoding ArsM, responsible for As methylation, thereby affecting the As methylation rate [ 20 ]. On the other hand, internal factors such as the expression of As metabolism genes (such as genes encoding As efflux transporters, arsB, etc.) may also affect microbial As methylation. However, little information is available about the genetic determinants of microorganisms that limit As methylation efficiency. Studies have shown that As methylation and volatilization in soil is a strictly biological process and driven by microbial activity [23][24][25]. Therefore, it is imperative to explore ways to accelerate the As volatilization rate of microorganisms for remediation of As-contaminated soil or water.
To date, most wild and engineered bacteria containing arsM often contain other genes encoding functions related to As resistance or transformation such as genes encoding As efflux transporters, arsA, arsD, arsB or acr 3, thus conferring two or more As detoxification mechanisms. For example, Rhodopseudomonas palustris strains (CGA009, TIE-1, HaA2 and BisB5) possess both arsM and genes encoding As efflux transporters (arsB or acr 3) [ 26 ]; Nostoc sp. PCC 7120 possesses both arsM and genes encoding As efflux transporters and/or C-As lyase (ArsI) [27]. As efflux transporters in the cell would reduce the intracellular As content, thereby conferring As resistance to microorganisms [28,29].
C-As lyase (ArsI) was shown to demethylate methylarsenite [MAs(III)] to As(III). Since both As(III) and MAs(III) can serve as substrates for ArsM, it is inferred that factors lowering the intracellular As(III) and MAs(III) content inevitably limit As methylation and subsequent As volatilization. More recently, studies have shown that demethylation of As limited the volatilization of As [30,31]. However, it is not known whether other internal factors that reduce intracellular As concentrations would limit microbial As volatilization.
Arsenic efflux is the most ubiquitous detoxification mechanism of microorganisms [ 32 ]. As efflux transporters reduce intracellular As concentrations by pumping out the corresponding As compounds, thus potentially limiting microbial As methylation and subsequent As volatilization. Previous studies have shown that deletion or inactivation of genes encoding As efflux transporters increased intracellular As accumulation [ 33 ]; Arsenicibacter rosenii SM-1 exhibited a higher As volatilization (47.6±18.4%) due to the lack of genes encoding As efflux transporters on its genome [9], and engineered E. coli strain lacking a gene encoding an As efflux pump displayed a higher As volatilization rate (~10.39%) [34]. Therefore, As efflux transporters limit the microbial As methylation rate.
R. palustris CGA009, a good model organism for studying As detoxification, has at least three ars operons (ars1, ars2 and ars3), which made it utilize different detoxification strategies under complex environments and the combined ars operons conferred a higher As resistance [ 26 ]. Qin et al. [ 8 ] identified and characterized ArsM from R. palustris CGA009 by heterologous expression in As sensitive E. coli AW3110(DE3) (∆arsRBC). Our previous study examined the expression of arsM from R. palustris CGA009 at the transcriptional level but conferred a relatively limited ability to volatilize As [ 26 ]. Moreover, the arsM derived from R. palustris CGA009 not only endowed the E. coli AW3110(DE3) (∆arsRBC) with As(III) resistance and volatilization rate (>10.39%), but also ArsM showed higher As volatilization capacity [ 34 ]. Curiously, the arsM from R. palustris CGA009 could be expressed and the encoded ArsM was active, but it conferred a relatively limited As volatilization rate for R. palustris CGA009. The reason may be that As efflux transporters pump As out of cells, resulting in the decrease of intracellular As concentration, which may restrict the process of As volatilization.
To test this hypothesis, we deleted arsB and acr3 from R. palustris CGA009, respectively, and obtained two mutant strains to investigate the effect of As efflux transporter on microbial As methylation. This work proposed an explanation for the low As volatilization of microorganisms, which provided an effective strategy for screening microorganisms with highly As volatilization, and laid a good foundation for bioremediation of Ascontaminated soil and water.

Results mutants
The suicide vector from E. coli WM3064 was transferred into R. palustris CGA009 by conjugation. Single colonies were picked on plates supplemented with 10% sucrose. These single colonies were considered to be potential mutants and confirmed by PCR. The genomic DNA extracted from the wild strain and from the mutants were used as template for PCR amplification with primers arsBU-F and arsBD-R ( Table 1). The amplified fragments of wild strains and mutants were 2500 bp and 2000 bp, respectively, which were consistent with the expected results, indicating that arsB deletion mutants were obtained bp for acr3 and that of the acr3 mutants was 1742 bp with primers acr3U-F and acr3D-R (Table 1), respectively ( Fig. 1b Lane 1-6). We were able to obtain two arsB mutants and six acr3 mutants, and then selected one strain for sequencing, respectively. The sequencing results showed that the deleted gene sequences were identical to that deposited in GenBank Database. The two mutants were named ∆arsB mutant and ∆acr3 mutant, respectively.

Determination of As resistance of mutants
To compare the ability of the wild type and the mutants to cope with As toxicity, the mutants and wild type were tested for As(III) and As(V) resistance (Fig. 2). Negligible difference was observed between the culture supplemented with 0.1 mmol/L of As(V) and the control (no As(V)), indicating that a low concentration of As(V) was not toxic to the wild type and the mutants. The growth rate of bacteria gradually decreased with increasing As(III) or As(V) concentrations. When As(III) and As(V) concentrations reached 3.0 mmol/L and 8.0 mmol/L, respectively, the growth of all strains was almost completely inhibited ( Fig. S1 of the Supplementary Material). The IC 50 of As(III) and As(V) for the wild type R. palustris CGA009 was 1.57±0.11 and 2.67±0.04 mmol/L, while that of R. palustris CGA009 ∆arsB mutant was 1.47±0.02 and 2.12±0.03 mmol/L, and that of R. palustris CGA009 ∆acr3 mutant was 1.21±0.07 and 1.76±0.12 mmol/L, respectively (Table S1 of the Supplementary Material). Compared to R. palustris CGA009, the mutants displayed lower resistance to As(III) and As(V), and the resistance to As(V) was significantly higher than that of As(III). These results also indicated that there was a difference in the As efflux activity of ArsB and Acr3, and the As efflux activity of Acr3 was higher.

As volatilization rate by mutants
To examine the As volatilization rate of the mutants in comparison to the wild type, the difference in total As content before and after bacterial growth was used to assess the As volatilization rate (Fig. 3). After incubation for 12 h and 24 h in the presence of 25.0 µmol/L As(III), the total As content in the medium decreased by 3.91% and 5.83% when incubated with the R. palustris CGA009 ∆arsB mutant, and that of R. palustris CGA009 ∆acr3 mutant decreased by 5.42% and 9.19%, respectively; whereas the total As content decreased by only 0.97% and 0.49% when incubated with the wild type R. palustris CGA009, respectively. When exposed to 100.0 µmol/L As(III) under the same conditions, the total As content in the medium decreased by 3.58% and 5.65% when incubated with the R. palustris CGA009 ∆arsB mutant, and that of R. palustris CGA009 ∆acr3 mutant decreased by 5.10% and 11.62%, respectively; whereas the total As content decreased by only 0.49% and 1.54% with the wild type R. palustris CGA009, respectively. Similarly, when exposed to As(V), the wild type and the two mutants also exhibited different As volatilization rates. After incubation for 24 h exposed to 100.0 µmol/L As(III), the R. palustris CGA009 ∆acr3 mutant showed the highest As volatilization rate with the rate reaching 12.40 % of the total As. It can be concluded that the efflux capacity of Acr3 is greater than that of ArsB, and the presence of both ArsB or Acr3 could significantly limit the As volatilization rate.

Arsenic speciation in mutants
Both wild type and mutants were exposed to 25.0 µmol/L As(III) for 24 h, the resulting As species were analyzed by HPLC-ICP-MS (Fig. 4). MAs(V) and DMAs(V) were detected, but their total accumulated concentration did not change significantly, while the As(III) concentration showed a major decrease. The residual As(III) in the medium was wild type R. palustris CGA009 > R. palustris CGA009 ∆arsB > R. palustris CGA009 ∆acr3, further indicating that the presence of As efflux transporters could significantly limit As volatilization rate.

Discussion
Arsenic contamination has raised worldwide concern due to the excessive anthropogenic release of As to the environment and its severe toxicity to organisms and ecosystems [1].
Therefore, there is an increasing interest in finding effective bioremediation tools able to remove As from As-contaminated environment. Currently, the concept of "green and sustainable remediation" has been advocated for heavy metal pollution in the world [35].
In view of the fact that bioremediation could better maintain soil structure and ecological balance, bioremediation of As by microorganisms has been advocated because of the microorganism's potential advantages in facilitating economically viable and environmental friendly technologies [36][37][38]. The final product of As methylation, gaseous trimethylarsine (TMAs(III)), could remove As from contaminated sites by As volatilization, thereby increasing the possibilities for bioremediation of As-contaminated environment [8,10,34]. As methylation is widespread in natural environment, and the genes encoding ArsM were shown to be distributed widely in various environments (e.g., paddy soil, groundwater, acid mine drainage and composting manure) [22,[39][40][41][42] and in every kingdoms of life [43]. However, the majority of microorganisms exhibited a relatively limited ability to volatilize As [25]. Besides environmental factors such as organic matter [13,18] , internal factors (such as genes encoding As efflux transporters, arsB, etc.) may also limit As volatilization rate of microorganisms. The As methylation rate not only determines whether As methylation is an important factor in detoxification [44], but also determines the application value of As methylation and subsequent volatilization.
Therefore, it is necessary to improve the As methylation efficiency of microorganisms for As contamination remediation.
Arsenic efflux is the most important As detoxification mechanisms of microorganisms. To date, six types of As efflux transporters (ArsB, Acr3, ArsJ, ArsP, MSF1, and ArsK) have been found in various As-resistant bacteria, and each was shown to translocate different types of As compounds out of the cell [32]. arsB and acr3 are both widespread genetic determinants encoding As(III) efflux transporters in As-resistant bacteria [ 45 ]. The MAs(III) efflux permease ArsP was identified in Campylobacter jejuni and shown to confer resistance to the organic arsenicals Rox(III) and MAs(III) but not to inorganic As(III) [ 46 ]. ArsJ and MSF1 were shown to extrude As(V) [ , not confer resistance to As(V) or dimethylarsenite [DMAs(III)] [ 49 ]. However, it is unclear whether the expression of these As efflux transporters would influence the microbial ArsM-mediated As methylation. Either arsB or acr3 is present in almost every prokaryotic species, and in some cases, both genes are present within a single organism, although no example of the coexistence of the two transporters encoded on the same operon has been reported [45]. ArsB proteins have only been detected in prokaryotes, whereas Acr3 proteins can be found in bacteria, archaea, fungi and some plants [45]. The arsB/acr3 encoded transporters pump As(III) out of the cell, which reduces the substrate concentration available for ArsM. Therefore, the presence of ArsB/Acr3 limits the As methylation efficiency of microorganisms. In this study, two As efflux genes (arsB/acr3) from R. palustris CGA009 were knocked out, respectively. The results showed that the resistance of the two mutants (R. palustris CGA009 ΔarsB and Δacr3) to As(III) and As(V) decreased, but the As volatilization rate increased. The R. palustris CGA009 Δacr3 mutant with a high As volatilization rate was obtained with the rate reaching 12.40 % of the total As (25.0 µmol/L As(III)). As efflux is a widespread and effective As detoxification pathway for microorganisms compared to As methylation [32], which was one of the possible reasons for the low efficiency of As methylation in R. palustris CGA009. These results indicated that As efflux transporters limited the As methylation efficiency of microorganisms. Compared to Arsenicibacter rosenii SM-1 with a higher As volatilization rate (47.6±18.4%) [9], the R. palustris CGA009 Δacr3 mutant exhibited a low As volatilization rate, which may indicate other reasons that limit As volatilization. On the one side, there is still a gene encoding ArsB in R. palustris CGA009; on the other side, low concentrations of As in the cells or low catalytic efficiency of ArsM will result in low As volatilization rate. Therefore, in a follow-up study, it may be possible to increase the As uptake by bacteria by knocking in genes such as glpF encoding the aquaglyceroporin channel GlpF or phosphate transporters, or to obtain higher activity of ArsM by increasing the number of copies of arsM or by directed evolution of ArsM, which may improve the As volatilization efficiency of the microorganism.
Previous studies had shown that ArsM encoded on the genome of R. palustris CGA009 displayed a high methylation acitivity [34]; therefore, a series of studies have been conducted with this enzyme, which is considered as a suitable candidate for an applied As methylation process in bioremediation [17,34,50,51]. However, this study and previous studies [26] showed that R. palustris CGA009 did not exhibit a significant As volatilization rate with the rate reaching less than 1% of the total As (25.0 µmol/L As(III)). Therefore, microorganisms with a highly active ArsM do not necessarily display a high As methylation rate, which may be due to internal factors (e.g., As efflux transporter) that limit the As volatilization rate. Past series of investigations have revealed As metabolism genes such as As efflux transporter genes or aio (arsenite oxidation) genes were commonly found in As-contaminated environments [41,[52][53][54]. These are all possible intrinsic factors that limit both As methylation and volatilization by lowering the intracellular As(III) content. In later studies, we intend to further investigate the effects of expression of other genes encoding functions related to As metabolism on microbial As methylation, and explore the factors that improve the As volatilization rate.

Conclusions
We demonstrated the interaction between As efflux transporters and ArsM by constructing mutants in genes encoding As efflux transporters. As efflux transporters were shown to be a remarkable intrinsic factor limiting As volatilization efficiency. The As volatilization efficiency was negatively related to As efflux activity of efflux transporters, and As volatilization rate could be significantly improved by deleting genes encoding microbial As efflux transporters. This work identified one of the reasons for the low efficiency of microbial As methylation, which could result in developing an effective strategies for modifying microorganisms to obtain higher As volatilization possibly creating an efficient bioremediation tool able to remove As from the environment.

Materials And Methods
Strains, plasmid, reagent and culture conditions Rhodopseudomonas palustris CGA009 (ATCC BAA-98) was obtained from the American Type Culture Collection (ATCC, USA) for anaerobic culture in modified Ormerod medium at 30°C with continuous illumination [55]. Escherichia coli strains were grown aerobically at 37°C in Luria-Bertani (LB) medium [56]. E. coli DH5α (Tiangen Biochemical Technology Co., Ltd., Beijing, China) was used for plasmid construction and replication. E. coli WM3064 served as the plasmid donor in conjugation with R.palustris, was a present by the Institute of Hydrobiology, Chinese Academy of Sciences. Plasmid pJQ200SK (Gm R ) was purchased from Miaolingbio, Inc. (Hubei, China). As(V) (Na 3 AsO 4 ·12H 2 O) and As(III) (NaAsO 2 ) were obtained from Merck (Darmstadt, Germany). The standard sample of As is As(III), As(V), methylarsonate [MAs(V)] and dimethylarsonate [DMAs(V)] mixed standard solution. The As concentration of each form is 10 μg/L, and As(V) and As(III) for standard preparation were purchased from Beijing Zhonglian Chemical Reagent Co., Ltd.; MAs(V) and DMAs(V) were purchased from AccuStandard. Inc (New Haven, CT, USA). All other used reagents were purchased from commercial sources, and were of analytical grade or better.

Construction and identification of R. palustris mutants
The arsB/acr3 gene of wild-type R. palustris CGA009 was deleted according to a previously described method [57]. The arsB gene (WP_011157811.1) on the ars1 operon was partially deleted, resulting in a frameshift mutation with gene knockout fragments of 500 bp (gene  (Table 1) and synthesized by Suzhou Genewiz Biotechnology Co., Ltd. (Suzhou, China). Primer design principles for amplification of upstream and downstream flanking DNA follow overlap extension PCR [58]. The upstream and downstream flanking DNA were amplified by PCR from R. palustris CGA009. All cloning methods, unless otherwise stated, were carried out as performed in Denman et al [56]. The DNA fragment and suicide plasmid pJQ200SK were double-digested with the corresponding restriction enzymes (Table 1)  As resistance assays R. palustris CGA009 and its mutant strain were anaerobically incubated until reaching an OD 660 of about 0.4 and inoculated in 4.0 mL modified Ormerod medium containing different As concentrations. Various concentrations of As(III) and As(V) were added with final concentrations ranging from 0.1 to 3.0 mmol/L and ranging from 0.1 to 8.0 mmol/L, respectively. Control experiments without As(III) were carried out under the same conditions. Each experiment was repeated in triplicate. Cell growth was estimated by measuring the OD 660 after incubation anaerobically at 30°C and with 2500 lux light for 4 days. The bacterial growth curve was plotted and the concentration (IC 50 ) at which As inhibited bacterial growth by 50% was calculated.
As volatilization of R. palustris mutants The difference in total As content before and after bacterial growth was used to assess the and DMAs(V) whose concentration of As in each form is 10 ug/L. The As was quantified by external calibration curves with peak areas. these works were performed at the Institute of Urban Environment, Chinese Academy of Sciences (Fujian, China).   The effect of As(III) (a) and As(V) (b) on growth inhibition rate of the mutants. R.
palustris CGA009 and its mutants were anaerobically grown at 30oC with continuous illumination. Error bars indicate the standard deviation from three independent experiments